Kroger nearly scraped the bottom of the barrel during the comparison, landing in the eighth place from the bottom of Sitemorse's list with a yawn-inspiring average response time of almost 1.5 seconds. How bad is that speed. "The general view of Sitemorse is that anything more than 250 milliseconds—a quarter of a second—is failing," a Sitemorse statement said.
Kroger was far from alone as an American E-Commerce site lacking get-up-and-go. Other sluggish sites were The Home Depot with a 570 millisecond response speed, Target at 616 milliseconds, Kohl's with 693 milliseconds, The Sports Authority at 739 milliseconds and Sears at a creaky 896 milliseconds.
A little bit better was Amazon with its 446 millisecond average speed, followed by Wal-Mart's average response time of 412 milliseconds, the 417 milliseconds of Macy's, Best Buy's 369 milliseconds and Apple's 364 milliseconds.
The Sitemorse system tested the top 125 pages of a site and averaged the time it took for those pages to react to standard "get" directives from a browser, Sitemorse Head of Client Services Gareth Evans said. Although 1.5 seconds might not seem particularly slow, it's a cyber-lifetime slower than the times attained by sites at the top of the Sitemorse list. Among the largest retail brands, J.C. Penney ranked the highest, coming in at number 40 with a spirited average response time of 141 milliseconds.
Sites owned by Walgreens, Dick's Sporting Goods, Limited Brands and Family Dollar were clustered relatively high on the list with response times between 224 milliseconds and 228 milliseconds. CVS, Foot Locker, Rite Aid and Dollar Tree also managed decent speeds, ranging between 234 milliseconds and 249 milliseconds.
Evans stressed that external factors, such as ISP and network speeds, have no bearing on Sitemorse's test results. Nor can a poor showing be blamed on one or two content-heavy pages in a site, such as a homepage full of Flash animation, he contended. "When we run our surveys, we have a look at the top 125 pages of those sites," he said. "We hit the homepage, check it with all our different tests, have a look for links on that page and follow them until we get 125 pages."
"We're just hitting all the pages we can find, as close to the homepage as we can. For each page we look at, we issue lots of 'get' requests and we track the response time of each of those requests across all of those 125 pages and then we average that," Evan said. "So it's not just the homepage or key landing pages. It's a real mixture of static pages and dynamic pages, and it's a pretty representative cross-section of all the pages in their sites."
As for what factors are likely to make one site faster than another, Evans said the differences are probably the result of server and database hardware quality or setup, although software does play a part.
"It could be some odd coding they've got in there or because they are using some creaky old technology," he said. "It could be they are having problems pulling data out of their database. Our figures are all about how long it takes for the hardware to respond to any request. In terms of what might be causing that, it could be a range of things. Maybe the hardware is not up to handling the volume of traffic due to design issues that don’t allow it to be speedy."
In fairness to those in the middle of the pack, Evans said a half-second average response time is decent. "Essentially, what we've got here is just a list of sites and those at the top, such as Debenhams (28 milliseconds), are very, very quick. Those are startlingly quick numbers, whereas Wal-Mart's speed is fair. It is still less than half a second. So we have to be a little bit careful of being critical of someone like Wal-Mart. We have to look at the numbers and say, 'It's half a second. Is that OK?' Half a second is alright, but speed is one of those things that are starting to come to the fore as connection speeds become quicker."